Hacker News new | past | comments | ask | show | jobs | submit
Gud | on: Progress on provable inference for large language models
If they are closed source and for profit, can they really be described as safety focused…